---
title: "OpenAI"
description: "Adds instrumentation for the OpenAI SDK."
url: https://docs.sentry.io/platforms/javascript/guides/nextjs/configuration/integrations/openai/
---

# OpenAI | Sentry for Next.js

For meta-framework applications running on both client and server, we recommend **setting up the integration manually** using the [`instrumentOpenAiClient` wrapper](https://docs.sentry.io/platforms/javascript/guides/nextjs/configuration/integrations/openai.md#manual-instrumentation) to ensure consistent instrumentation across all runtimes.

## [Automatic Instrumentation](https://docs.sentry.io/platforms/javascript/guides/nextjs/configuration/integrations/openai.md#automatic-instrumentation)

*Import name: `Sentry.openAIIntegration`*

If you are using a different runtime (like Bun, Cloudflare Workers or a Browser) or experiencing missing spans, you need to use **[Manual Instrumentation](https://docs.sentry.io/platforms/javascript/guides/nextjs/configuration/integrations/openai.md#manual-instrumentation)** to explicitly wrap your AI client instance instead.

The `openAIIntegration` adds instrumentation for the [`openai`](https://www.npmjs.com/package/openai) SDK to capture spans by wrapping OpenAI SDK calls and recording LLM interactions.

In Node.js runtimes, this integration is enabled by default and automatically captures spans for OpenAI SDK calls (requires Sentry SDK version `10.28.0` or higher).

To customize what data is captured (such as inputs and outputs), see the [Options](https://docs.sentry.io/platforms/javascript/guides/nextjs/configuration/integrations/openai.md#options) in the Configuration section.

## [Manual Instrumentation](https://docs.sentry.io/platforms/javascript/guides/nextjs/configuration/integrations/openai.md#manual-instrumentation)

*Import name: `Sentry.instrumentOpenAiClient`*

The `instrumentOpenAiClient` helper adds instrumentation for the [`openai`](https://www.npmjs.com/package/openai) SDK to capture spans by wrapping OpenAI SDK calls and recording LLM interactions with configurable input/output recording. You need to manually wrap your OpenAI client instance with this helper.

See example below:

```javascript
import OpenAI from "openai";

const openai = new OpenAI({
  // Warning: API key will be exposed in browser!
  apiKey: "your-api-key",
});

const client = Sentry.instrumentOpenAiClient(openai, {
  recordInputs: true,
  recordOutputs: true,
});

// Use the wrapped client instead of the original openai instance
const response = await client.chat.completions.create({
  model: "gpt-4o",
  messages: [{ role: "user", content: "Hello!" }],
});
```

To customize what data is captured (such as inputs and outputs), see the [Options](https://docs.sentry.io/platforms/javascript/guides/nextjs/configuration/integrations/openai.md#options) in the Configuration section.

## [Configuration](https://docs.sentry.io/platforms/javascript/guides/nextjs/configuration/integrations/openai.md#configuration)

### [Options](https://docs.sentry.io/platforms/javascript/guides/nextjs/configuration/integrations/openai.md#options)

The following options control what data is captured from OpenAI SDK calls:

#### [`recordInputs`](https://docs.sentry.io/platforms/javascript/guides/nextjs/configuration/integrations/openai.md#recordinputs)

*Type: `boolean` (optional)*

Records inputs to OpenAI SDK calls (such as prompts and messages).

Defaults to `true` if `sendDefaultPii` is `true`.

#### [`recordOutputs`](https://docs.sentry.io/platforms/javascript/guides/nextjs/configuration/integrations/openai.md#recordoutputs)

*Type: `boolean` (optional)*

Records outputs from OpenAI SDK calls (such as generated text and responses).

Defaults to `true` if `sendDefaultPii` is `true`.

**Usage**

Using the `openAIIntegration` integration for **automatic instrumentation**:

```javascript
Sentry.init({
  dsn: "____PUBLIC_DSN____",
  // Tracing must be enabled for agent monitoring to work
  tracesSampleRate: 1.0,
  integrations: [
    Sentry.openAIIntegration({
      // your options here
    }),
  ],
});
```

Using the `instrumentOpenAiClient` wrapper for **manual instrumentation**:

```javascript
const client = Sentry.instrumentOpenAiClient(openai, {
  // your options here
});
```

## [Supported Operations](https://docs.sentry.io/platforms/javascript/guides/nextjs/configuration/integrations/openai.md#supported-operations)

By default, tracing support is added to the following OpenAI SDK calls:

* `chat.completions.create()` - Chat completion requests
* `responses.create()` - Response SDK requests

Streaming and non-streaming requests are automatically detected and handled appropriately.

When using OpenAI's streaming API, you must also pass `stream_options: { include_usage: true }` to receive token usage data. Without this option, OpenAI does not include `prompt_tokens` or `completion_tokens` in streamed responses, and Sentry will be unable to capture `gen_ai.usage.input_tokens` / `gen_ai.usage.output_tokens` on the resulting span. This is an OpenAI API behavior, not a Sentry limitation. See [OpenAI docs on stream options](https://platform.openai.com/docs/api-reference/chat/create#chat-create-stream_options).

## [Supported Versions](https://docs.sentry.io/platforms/javascript/guides/nextjs/configuration/integrations/openai.md#supported-versions)

* `openai`: `>=4.0.0 <7`